Thursday, December 14, 2017
Search
  
Submit your own News for
inclusion in our Site.
Click here...
Breaking News
T-Mobile Enters The pay-TV Market With the Acquisition of Layer3 TV
Samsung Patents 'Double-sided' Smartphone
Apple Invests in FaceID Component Maker Finisar
Corsair Launches 4GHz DDR4 SODIMM 32GB Memory Kit
Google's Year in Search
Google Opening Artificial Intelligence Research Center in China
AT&T is Testing Broadband Internet Over Power Lines
Toyota and Panasonic to Start Study on Automotive Prismatic Battery Business
Active Discussions
Which of these DVD media are the best, most durable?
How to back up a PS2 DL game
Copy a protected DVD?
roxio issues with xp pro
Help make DVDInfoPro better with dvdinfomantis!!!
menu making
Optiarc AD-7260S review
cdrw trouble
 Home > News > General Computing > Machine...
Last 7 Days News : SU MO TU WE TH FR SA All News

Thursday, August 03, 2017
Machine Learning Could Identify What You Are Typing On Skype By Recognizing Your Keyboard's Clicks


Many of us talk on Skype, Hangouts, WhatsApp, or Viber while using the computer for something else. But have in mind that your conversation partner might find out what you are typing.

Click, click, click - almost every person talking on the phone or Skype knows that his or her conversation partner may be chatting or doing something else while conversing. The sound of typing on a physical computer keyboard is rather recognizable.

With help from machine learning and a computer it is possible to find out what your conversation partner is typing. Almost every keyboard has a slightly different sound for each key, which means you can use a recording of typing to find out which keys have been pressed. The accuracy is not perfect, but it is quite high nonetheless.

At the Black Hat hacking conference in Las Vegas, presentation showed how to make it happen. According to the researchers, even after the sound has undergone conversions during online transmission, the recordings of key clicks retain sufficient information to feed to a machine-learning system and get back the five most probable key presses. The results for the three test laptops that researchers used to conduct this study were different, but it was possible to retrieve the typed text even for the most "indistinctly clicking" Lenovo laptop.

This is where dictionary-based results come in. People generally type words that make sense; thus, any nonsense can be eliminated. It is enough to know which keyboard layout and language the victim probably used; the demonstration algorithm will deliver an intelligible and plausible result within a few seconds.

The experts claim that this technology might be used even to steal passwords, although that sounds a bit far-fetched. Passwords are too short and do not consist of real words most of the time.

The threat of "input interception" over Skype does not look too serious, but it's worth knowing about, especially if you sometimes deal with confidential information.

However, if you find yourself in a middle of an especially long and tiresome conference call, observe the golden rule: Anyone who is not speaking should mute his or her microphone until it's time to talk.



Previous
Next
Razer Wants to Take Chroma Lighting Beyond Peripherals        All News        Apple Leads the Declining Tablet Market
Google is Developing technology for Snapchat-like media Content     General Computing News      Facebook to Take New Fake News Measures

Get RSS feed Easy Print E-Mail this Message

Related News
IBM Scientists Demonstrate 10x Faster Machine Learning Using GPUs
Google Detects Diabetic Eye Disease With Machine Learning
Google DeepMind Introduces Method for Improving Machine Learning Speed
Google Enhances Translation, Your Old Photos And Goes All In on Cloud Machine Learning
Google Opens Machine Learning Tools To Cloud Customers
Google Used Machine Learning To Master Game of Go
Scientists Teach Computers Learn Like Humans

Most Popular News
 
Home | News | All News | Reviews | Articles | Guides | Download | Expert Area | Forum | Site Info
Site best viewed at 1024x768+ - CDRINFO.COM 1998-2017 - All rights reserved -
Privacy policy - Contact Us .